To achieve fast location, precise tracking and accurate identification over a large field of view (FOV), we have proposed a heterogeneous compound eye camera (HeCECam), which consists of a heterogeneous compound eye array, an optical relay system and a CMOS detector. However, the current HeCECam can hardly acquire high-precision 3D information of the targets to realize these applications. To solve this challenge, we propose a scheme on optimizing the structure of the HeCECam to improving the detection performance, including the optimization of the distribution uniformity of the sub-eyes with the proposed "Three-direction center-of-gravity subdivision (TGS)" and the enhancement of the compatibility between heterogeneous compound eyes and the optical relay system with the proposed compensation method for tilt. The TGS significantly reduces the distribution unevenness of sub-eyes down to 117% from the previous 152%, and provides symmetry to the heterogeneous compound eye array. The tilt compensation effectively addresses previous imaging defects, such as distortion of sub-images, increased stray light, and support structures being imaged, and it improves the imaging clarity of the system, especially in external FOV. Based on two proposed methods, we re-design and fabricate the heterogeneous compound eye array to obtain a high-performance prototype. To verify the imaging capacities of the optimized HeCECam, a series of comparison experiments are performed, including blank scene imaging, FOV tests, resolution verification and real-world scene imaging. The results show that the previous imaging defects have been well eliminated, and the optimized prototype has stronger resolving power and wider FOV. This allow the HeCECam to perform better in subsequent practical applications, such as wide-area surveillance, forewarning, and navigation.